In 1807, Benjamin Waterhouse was caught in a state of curatorial embarrassment, setting the stage for his dismissal from Harvard. The trouble had started five years earlier, when Samuel Webber, the Hollis Professor of Mathematics, requested that Waterhouse remove the mineral cabinet from the Philosophy Chamber of Harvard Hall because the expanding collection interfered with Webber’s lectures. Over the short-term, the preeminent Waterhouse successfully deflected Webber’s complaints. But in 1806, Harvard’s Board of Overseers tapped Webber to serve as the new president of the College. Early the next year, Webber and his friend Judge John Davis retaliated against Waterhouse; they made a “surprise” inspection of Waterhouse’s mineral cabinet.
They were greeted by utter chaos. Six or seven hundred of the over 1,600 specimens were without labels, and no memoranda indicated their identities. An incomplete attempt to classify the minerals disrupted their initial arrangement within the drawers, and sixty-eight specimens were missing altogether. Waterhouse produced all but three of these specimens, but the members of the Harvard Corporation, the university’s smaller governing body, were not appeased. They issued Waterhouse a stern warning to separate “the foreign from the domestic specimens” and to develop a mineral catalogue as soon as was practicable. If such a catalogue materialized, there is no record of it. On November 6, 1809, the Corporation formally relieved Waterhouse of his duties as medical professor.
The mess that Waterhouse made of the Harvard mineral collection—which today numbers over 100,000 specimens and is both the oldest and the largest university mineral museum in the United States—illustrates a significant but little-known crisis that raged within the field of mineral sciences at the turn of the nineteenth century. In the early 1800s, there was no simple way to identify a mineral. When minerals within a reference collection were separated from their labels, the restoration of order was nearly impossible without a trained specialist. Even when a mineral’s identity could be discovered, natural historians could not agree on a universal system by which minerals should be organized. Indeed, they argued over the question of taxonomic relativism: Are minerals best grouped through a single classification scheme that captures “natural truth”? Or, can minerals be categorized by a multiverse of equally valid but philosophically distinct rules?
In this respect, mineralogical disciples of natural philosophy gazed jealously at their biological counterparts, whose world was almost miraculously well charted by the organizational system developed in the mid-1700s by the Swedish botanist and zoologist, Carl von Linné (1707-1778). Linnaean classification is based upon a binomial logic that sorts organisms into bins of ever increasing specificity, and the criteria by which organisms are evaluated are macroscopic and quantifiable. Any animal or plant can be categorized into a hierarchical sequence of domains, kingdoms, phyla, classes, orders, families, genera, and species. Linnaeus himself believed that he had cracked the Divine Code, and in an autobiography, he proclaimed, “No one has more completely changed a whole science and initiated a new epoch.”
Linné did not restrict the application of his organizational system to the living world. Linné regarded rocks and minerals as the products of Creation along with life, and he extended his taxonomy to stones. He classified minerals within nested bins of classes and orders, and he applied to minerals the same binomial nomenclature by which we are labeled as Homo sapiens (for example, Quartzum aqueum for clear quartz, Quartzum album for white quartz, and Quartzum tinctum for colored quartz).
[media:node/3114 caption vertical medium left] The Linnaean taxonomy for minerals—which came to be known as the Natural System or the Natural-History System of mineral classification—predominated from the mid-eighteenth to the mid-nineteenth century. Its most famous exponent was Abraham Gottlob Werner, a professor at the mining academy in Freiberg, Germany, who wrote the enormously influential Treatise on the External Characteristics of Fossils in 1774, four years before Linné’s death. The term “fossil” was used at that time to denote any material found within the Earth, and most typically it designated minerals rather than petrified organisms. Werner was determined to characterize the physical attributes of minerals with the same rigor with which Linnaean biologists described plants and animals. HisTreatise specifies 77 varieties of color, with red alone apportioned into 15 types: blood-red, flesh-red, scarlet-red, cherry–red, and so forth. He was similarly exacting in his descriptions of mineral shape, surface texture, luster, and cleavage.
In this way, proponents of the Natural-History System were aiming to be scientific in their efforts to contextualize minerals within a single philosophical framework. Unlike organisms, however, which exhibit a preponderance of discrete body parts and behaviors that enable classification with precision, the physical characteristics of minerals can be shaped by growth environments, chemical impurities, and post-crystallization processes. These can render two specimens of the same species macroscopically different, or two specimens of different species apparently similar.
Contemporaries of Linné and Werner recognized the deficiencies of this approach, and they advocated for a mineral classification system that would function more universally. A small coterie of scholars argued that physical characteristics are subordinate to a more defining attribute—chemical composition. Johan Gottschalk Wallerius (1709–1785) and Axel Fredrik Cronstedt (1722–1765) of Sweden, the Russian polymath Mikhail Vasilyevich Lomonosov (1711–1765), and the Edinburgh Professor of Natural History John Walker (1731–1803) contended that the major elements that constitute minerals provide the proper basis for sorting them. These visionary champions of the Artificial System laid the groundwork for the mineral classification scheme that ultimately would prevail, but they faced stiff opposition during their lifetimes. After all, only about 40 elements had been identified by the turn of the eighteenth century, and the techniques for measuring mineral composition were primitive. Indeed, one can make a case that, in the mid-1700s, the acute description of macroscopic mineral characteristics was a more scientific approach than the distillation of minerals into their chemical components.
The tide started to shift when the famed French chemist, Antoine Lavoisier (1743–1794), redefined our notion of “elements” and debunked phlogiston-based theories. The quantitative rigor with which Lavoisier endowed his investigations inspired others to follow suit, and by 1800 the Artificial system of ordering minerals was gaining traction.
In the following year, however, a new twist to the debate began when a French priest published a five-volume Traité de Minéralogie. As an ardent botanist, René Just Haüy (1743–1822) admired the symmetry exhibited by flower petals. In an oft-told story, Haüy accidentally dropped a crystal of calcite loaned to him by a friend, and he was fascinated by the geometric uniformity of the cleavage fragments. Haüy proceeded to develop the hypothesis that minerals are composed of fundamental particles that he called molécules intégrantes. When the angles between the faces of these particles and the ratios of the edge lengths are measured precisely, Haüy averred that every mineral could be uniquely pigeon-holed within his novel Geometric System. Haüy immediately won over natural historians who saw in his approach the ideal resolution to the battle between mineral classification stratagems.
Despite the ardor of its advocates, the Geometric system also drew its critics. Some minerals with very different characteristics fracture into fragments that exhibit identical shapes. For example, common table salt, known as halite to mineralogists, and a common ore of lead called galena both will cleave into perfect cubes, but they are otherwise dissimilar. Common salt crystals are transparent, glassy, and light-weight, whereas galena is dense and metallic. This mineral pair exemplifies a phenomenon known as isomorphism: they have different chemical compositions but—as we know today—identical atomic architec-tures. They are like two houses constructed according to the same blueprint but erected from different building materials. The German chemist Eilhard Mitscherlich (1794-1863) proved the reality of isomorphism in 1819, and he took a dim view of the Geometric school and of the proliferation of classification systems. In an 1824 letter, he wrote, “Everyone…is developing a system of mineralogy of his own, according to his own method; I do not expect much good will result from this.”
[media:node/3112 caption vertical medium left] The mineralogy community was in desperate need of an intellectual giant who could cut this Gordian knot and place mineral classification on the rigorous footing with which Linné had ordered living things a century earlier. And indeed, during the year of Mitscherlich’s letter, a system materialized with the cure for this mineralogical dyspepsia. Nearly every college Earth science text identifies the father of this system as the prolific Professor of Natural History and Geology at Yale from 1850 to 1892—James Dwight Dana. Dana (1813–1895) published his first System of Mineralogy at the age of 24 while under the tutelage of the country’s most illustrious scientist, Benjamin Silliman. Dana published five editions of this highly influential work during his lifetime. The eighth edition was published in 1997.
Ironically, however, James Dwight Dana was not the genius who conceived the solution to the controversy. Dana’s role in this drama is more akin to that of Galileo in the triumph of the heliocentric model of the solar system; he was more popularizer than pioneer. The Copernicus of our story is the recipient of Mitscherlich’s plaintive letter: a Swedish scientist named Jöns Jacob Berzelius (1779–1848), who stands unfairly shadowed behind Lavoisier in the invention of modern chemistry. Berzelius devised the system by which elements are designated by symbols: H for hydrogen, O for oxygen. He also took the next step and created the molecular formula: H2O, though he used superscripts (H2O) rather than subscripts and often represented oxygen atoms by dots. More significantly, Berzelius analyzed thousands of minerals to show that mineral compounds consist of groups of atoms in constant proportions, as hypothesized by the English chemist John Dalton (1766–1844).
To prove these ideas, Berzelius exploited a precursor to the modern battery called the voltaic pile. He observed that minerals dropped into the electrolyte solution of the battery decompose when a voltage is applied. Certain elements, such as oxygen, consistently migrate to and react with the positively charged electrode. Other elements, such as metals, tend to plate the negatively charged electrode. In this way, Berzelius laid the cornerstone for what today we call electrochemical dualism—the notion that most inorganic compounds are the union of oppositely charged elements. As an avid mineral collector, Berzelius immediately grasped the potential for his model to offer a new basis for mineral classification. But he faced a dilemma: Should mineral taxonomy be founded on the positive or the negative elements in mineral compounds? Berzelius opted to go positive. In 1814, he offered a system that organized minerals according to their shared metallic elements.
The adherents of the Natural school pounced. Thomas Thomson (1773–1852), a Scottish chemist who founded the Wernerian Natural History Society of Edinburgh, pointed out “that these definite proportions, this chemical composition according to the atomic theory, can be perceived only in a small number of individuals.” Impurities in natural minerals and the absence of standardized methods for chemical analyses yielded a high level of variation in the formulas derived for a given species. Thus, chemistry at the time was arguably less “scientific” than an acute visual characterization of the observable features of a mineral specimen.
Even Berzelius had to admit that his system yielded some illogical outcomes, and in 1824 he revised his approach to group minerals according to their negatively charged chemical component. Here at last was a system that united calcite, magnesite, siderite, and rhodochrosite into the same clan because they share the so-called carbonate group (CO32-). The carbonate radical (a chemical term for charged molecular groups invented by Berzelius, along with his coinages of polymer and catalysis) dominantly endows these minerals with their physical and chemical behaviors. Similarly, Berzelius grouped together minerals containing oxygen as the only negative entity into oxides, sulfur as the negative component into sulfides, and silicon and oxygen radicals into silicates.[media:node/3113 caption vertical large right]
In short, Berzelius was the first to intuit the philosophical germ of the taxonomy that mineralogists use today. Unfortunately for Berzelius, his contemporaries treated the revised taxonomy with the disdain that greets politicians who flip-flop on major policy issues. Even his own students were divided in their adoption of his first and second systems.
In modern mineralogical mythology, it was at this stage that James Dwight Dana divined the grand solution through his ground-breaking System of Mineralogy in 1837. Admittedly, Dana’s first System received positive reviews in the United States and abroad, but contrary to most perceptions, it did not instigate a revolution. The 1837 edition of the System adopted a traditional Natural-History framework straight out of Linnaeus, from the organization of minerals into orders and genera to the use of a Latinate binomial nomenclature. Dana labeled diamond, for example, as the species Adamas octahedrus. More objectionably, he classified diamond alongside quartz, sapphire and beryl in the Order Hyalinea, because he judged the translucent glassiness of these minerals to trump chemical composition as a measure of mineral affinity.
As the craft of chemical analysis became increasingly refined towards the middle of the nineteenth century, however, the evidence for “these definite proportions, this chemical composition according to the atomic theory” became overwhelming. Dana could no longer deny that the negatively charged chemical component in the min-eral was the best predictor of min-eral affinities—as Berzelius had ultimately argued. In his fourth (1854) edition of the System, Dana offered his unqualified surrender. He presented a taxonomy “in which the Berzelian method was coupled with crystallography.” Dana divided the mineral kingdom into primary groups that included Native Elements, Sulfides, Halides, and Oxides, and the Oxides were subdivided into Silicates, Phosphates, Sulfates, Carbonates, and these groups were further subdi-vided based on their electronegative chemistry.
[media:node/3111 caption vertical medium left] The acceptance of the Dana mineralogical system achieved closure after a century of acrimonious debate. It fundamentally transformed the way that scientists view the stuff of our world, and it required a leap of faith. The scientists who rejected the Natural-History School of mineral classification were forced to abandon the physical characters that they could see, touch, taste, and smell. In the middle of the nineteenth century—when only half of all elements were known—and twenty-five years before Mendeleev developed the Periodic Table—and sixty years before crystal structures were first revealed by X-ray diffraction—mineralogists agreed to sort minerals based on what was then the ultimate unknown—the atomic building blocks of all matter.
It is striking that Dana judged it necessary to divest his mineral classification system of all trappings of a Linnaean hierarchy. Instead, following Dana’s rules, minerals are named by a single word ending in –ite, except for the handful that are grandfathered, such as quartz and topaz. In retrospect, Dana’s elimination of the hierarchical classes, orders, and families, and his abdication of a binomial nomenclature, were entirely unnecessary. Like the Linnaean taxonomy for organisms, Dana’s logic sorts minerals by a binary algorithm, and the Dana system classifies minerals within a series of ever more chemically constrained bins. So, minerals can be mapped on a Tree of Minerals that is analogous to the Tree of Life. For example, the mineral hematite, which is a main component of rust and occurs within the family of minerals having the corundum crystal structure, could be called Corundum hematitus. But Dana’s complete rejection of the Natural School’s symbology was integral to his transition to a new paradigm.
[media:node/3110 caption vertical large right] By 1870, the Dana classification system for minerals was adopted nearly universally. But what is the rationale for its success? Why does a chemically based Berzelian approach work so well for sorting minerals? My own discussions with colleagues reveal a sense that the Dana system triumphed because it is the least ambiguous, and thus the most reproducible and rigorous, of the possible organizing methods. But the Dana system in fact is rife with ambiguity. As Dana, himself knew, many minerals exhibit multiple chemical constituents that are negatively charged. The copper-based minerals azurite and malachite, used historically as blue and green pigments in Renaissance paintings, each contain negatively charged carbonate and hydroxyl entities. Do we group them with carbonate or hydroxide minerals? (Dana arbitrarily identified them as carbonates.) The International Mineralogical Association exists largely to oversee the classification of new (and sometimes old) minerals precisely because the task is so arduous. The gem material tourmaline, for example, is symbolized by the formula Na(Mg,Fe,Li,Al,Mn)3Al6(BO3)3 Si6O18(OH,F)4.
If scientific rigor does not explain why Dana’s system seems so “correct”, then what does? Surprisingly, the answer has become apparent only within the last decade, thanks to the insights of a modern natural philosopher, Robert Hazen, of the Carnegie Institution of Washington. The Carnegie’s Geophysical Laboratory is a geologist’s think-tank with an illustrious history of discovery. Since 2008, Hazen and his collaborators have published a series of articles promoting their thesis that minerals have evolved over time. Mineral evolution is not precisely analogous to organic evolution. We do not presume that the minerals that compose Earth’s present landscape emerged through the coupled extinction and genealogical adaptation of a long series of precursors, as is the case with life forms. But the Danan Tree of Minerals is endowed with time.
Over the last 50 years, geologists have demonstrated that our Earth developed through a series of episodic transformations. These include the separation of accreted planetesimals into an interior iron core and a magnesium silicate mantle almost immediately after the Earth’s assembly 4.54 billion years ago. Millions of years later, the Earth resurfaced itself with aluminosilicate basalts through intensive volcanism, followed by the partial melting of those basalts to make the more siliceous rocks that would ultimately create continents. Perhaps 3.5 billion years ago, a global overturning of the mantle initiated plate tectonics, and a billion years after that, Earth’s atmosphere and oceans saw a rise in oxygen due to the proliferation of bacterial photosynthesis.
These sporadic reinventions of our planet involved global changes in its chemistry. The record of these events is preserved in Earth’s mineralogy, which acts as an indicator for chemical transformation as surely as fossils of once-living things provide a record of biological evolution. Like organisms, minerals are subject to a kind of natural selection, because as new geologic ecosystems were created, some minerals reacted to form new species that were better “adapted”—which in this context usually means more energetically stable—than their precursors. Hazen and colleagues are working to quantify the diversification of Earth’s mineralogy through time, from the dozen or so residual “ur-minerals” from the condensing solar nebula to the many thousands that inhabit the planet today. The appearance of new mineral families involved periods of stasis followed by intense diversification as new global processes were initiated.
And that, in the end, is why Berzelius—and his reluctant apostle, Dana—were right. The Dana classification system contains the code for Earth’s own history. Entire branches of Dana’s Tree of Minerals awaited the proper combination of temperatures, pressures, and chemical ingredients before they could make their appearance on our planet. Metal sulfides, for example, likely dominated crustal mineralogy until the oxidation of Earth’s atmosphere converted them to oxides, hydroxides, and hydrates. Hazen argues that fully two-thirds of Earth’s mineralogy awaited this preparatory event.
It took one hundred years after Linné first offered his classification scheme before Darwin intuited his theory of natural selection, and it is inconceivable that he could have made his intellectual leap without a Linnaean Tree of Life to guide his thinking. Similarly, geologists developed their understanding of planetary differentiation, plate tectonics, and the Great Oxidation Event one hundred years after the acceptance of Dana’s classification system. Though generally unacknowledged, those geologic breakthroughs never would have transpired without Dana’s mineral map to direct them. --PJH